Human Vital Signs Parameters Estimation Using Millimeter Wave Radar

The electromagnetic wave signals transmitted by radar detect the displacement of the human chest wall caused by the physiological activities of respiration and heartbeat,to achieve the accurate estimation of vital sign parameters,which has important research value and application prospects in the fields of clinical medical care,patient monitoring,smart home,fatigue driving detection,etc.In this thesis,the research of the human vital sign parameter estimation method is carried out based on a millimeter wave radar platform,including human vital sign signal modeling,target detection and localization,multi-target signal separation,single-target vital sign parameter estimation,etc.The main work is as follows:1.A typical human vital sign signal model was established.By analyzing the regular pattern of human respiration and heartbeat physiological activities and combining the features of frequency-modulated continuous wave radar,the intrinsic connection between radar echo and human chest vibration is characterized,and the method of extracting the phase sequence containing human vital sign information from radar echo data is elaborated,and its effectiveness is verified by simulation experiments.2.In the single-target scenario,parameter estimation methods applicable to different measurement time lengths are researched for problems such as weak heartbeat amplitude,low signal-to-noise ratio,and insufficient robustness for long-term monitoring.For longterm monitoring,the method based on correntropy spectral density is researched,and the accuracy of heart rate detection is effectively improved by using the superiority of correntropy spectral density in the non-Gaussian background;the correction algorithm based on historical value assisted is proposed,and the current measurement value is estimated by using historical results assisted,which effectively improves the robustness of the algorithm in long-term monitoring.For short-time measurement,an improved HHT-based vital sign parameter estimation method is researched,which makes use of the high resolution of Hilbert Huang transform to compensate for the lack of resolution of the spectral estimation algorithm in short-time measurement,and also uses the variational mode decomposition instead of empirical mode decomposition,which effectively improves the noise-resistance performance of the algorithm.3.In the multi-target scenario,the multi-dimensional joint signal separation method of range,angle,and source is researched for the problem of mutual coupling interference of targets,and three methods of range-bin data extraction,beamforming,and blind signal separation are used for different position relationships between targets,respectively,which effectively separate the multi-target signals and achieve multi-target vital sign parameter estimation.The above algorithms are validated by the actual measurement data,which proves the effectiveness of the algorithms and provides strong theoretical and technical support for the engineering application of human vital sign parameter estimation.